Available online http://breast-cancer-research.com/content/7/6/R890

Open AccessVol 7 No 6Research articleCYP17 gene polymorphism in relation to breast cancer risk: a case-control studyKristjana Einarsdóttir1,2, Tove Rylander-Rudqvist3, Keith Humphreys1, Susanne Ahlberg4, Gudrun Jonasdottir1, Elisabete Weiderpass1,5, Kee Seng Chia6, Magnus Ingelman-Sundberg4, Ingemar Persson1,7, Jianjun Liu2, Per Hall1 and Sara Wedrén1,2,6

1Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden2Population Genetics, Genome Institute of Singapore, Singapore3Division of Nutritional Epidemiology, Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden4Division of Molecular Toxicology, Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden5Cancer Registry of Norway, Montebello, Oslo, Norway6Centre for Molecular Epidemiology, Department of Community, Occupational and Family Medicine, National University of Singapore, Singapore7Swedish Medical Products Agency, Uppsala, Sweden

Corresponding author: Kristjana Einarsdóttir, Kristjana.Einarsdottir@meb.ki.se

Received: 25 May 2005 Revisions requested: 8 Jul 2005 Revisions received: 15 Aug 2005 Accepted: 18 Aug 2005 Published: 14 Sep 2005

Breast Cancer Research 2005, 7:R890-R896 (DOI 10.1186/bcr1319)See related Commentary: http://breast-cancer-research.com/content/7/6/238This article is online at: http://breast-cancer-research.com/content/7/6/R890© 2005 Einarsdóttir et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Introduction The c.1-34T>C 5' promoter region polymorphismin cytochrome P450c17 (CYP17), a key enzyme in thebiosynthesis of estrogen, has been associated with breastcancer risk, but most previous studies have been relatively small.

Methods We genotyped 1,544 incident cases of primary breastcancer and 1,502 population controls, all postmenopausalSwedish women, for the CYP17 c.1-34T>C polymorphism andcalculated odds ratios (ORs) and 95% confidence intervals(CIs) from logistic regression models.

Results No overall association was found between CYP17 c.1-34T>C and breast cancer risk, OR 1.0 (95% CI 0.8–1.3) for the

A2/A2 (CC) carriers compared to the A1/A1 (TT) carriers,regardless of histopathology. We detected an interactionbetween CYP17 c.1-34T>C and age at menarche (P = 0.026)but regarded that as a chance finding as no dose-responsepattern was evident. Other breast cancer risk factors, includingmenopausal hormone use and diabetes mellitus, did not modifythe overall results.

Conclusion It is unlikely that CYP17 c.1-34T>C has a role inbreast cancer etiology, overall or in combination withestablished non-genetic breast cancer risk factors.

IntroductionEstrogen is a central factor in the etiology of breast cancer [1].One of the key enzymes in the synthesis of sex hormones,such as estrogens and androgens, is cytochrome P450c17(CYP17). CYP17 catalyses the 17α-hydroxylation of pregne-nolone and progesterone, and these intermediates are thenconverted to dehydroepiandosterone and androstenedione bythe 17,20-lyase activity of the enzyme [2,3].

A single T (A1 allele) to C (A2 allele) base change in the 5' pro-moter region of CYP17 (c.1-34T>C) has been suggested to

create an additional binding site for the transcription factorSp1 [4]. This could theoretically lead to increased levels of theenzyme, but use of the extra binding site has not been con-firmed experimentally [5]. Three groups have reported anassociation of the A2 allele with increased breast cancer risk[6-8] but others have failed to replicate those findings [5,9-21]. A recent meta-analysis of 15 case-control studies did notfind any overall association [22] but this analysis was criticizedby Feigelson and colleagues [23], who found a borderline sig-nificant association between the CYP17 polymorphism andadvanced breast cancer.

R890CI = confidence interval; CYP17 = cytochrome P450c17; DASH = dynamic allele-specific hybridization; OR = odds ratio; PCR = polymerase chain reaction; RFLP = restriction fragment length polymorphism.

Breast Cancer Research Vol 7 No 6 Einarsdóttir et al.

R891

In a population-based study about causes of postmenopausalbreast cancer, which included 1,544 cases and 1,502 con-trols, we evaluated the association between CYP17 c.1-34T>C and breast cancer risk among Swedish women. Wealso explored potential interaction between genotype andmenopausal hormone use and diabetes mellitus as well as ageat menarche, age at menopause, age at first birth, parity andbody mass index.

Materials and methodsFounding studyThis nation-wide case-control study included all Swedishwomen, 50 to 74 years of age and resident in Swedenbetween October 1993 and March 1995. During that period,all incident cases of primary invasive breast cancer were iden-tified at diagnosis through the six regional cancer registries inSweden, to which reporting of all malignant tumors is manda-tory. Controls that matched the cases in 5-year age stratawere randomly selected from the Swedish Registry of TotalPopulation. Of the eligible cases and controls, 3,345 (84%)and 3,454 (82%), respectively, participated in the study byproviding detailed information via questionnaires about meno-pausal hormone use, reproductive history, and other lifestylefactors. Results from the founding study have been published[24-28].

Collection of biological samplesOnly postmenopausal women with a first primary breast can-cer without any previous malignancy (except carcinoma in situof the cervix or non-melanoma skin cancer) were eligible forthe present study. Premenopausal women and those with aprevious malignancy were excluded. From the eligible partici-pants, we randomly selected 1,500 cases and 1,500 age-fre-quency (in 5-year groups) matched controls. We wereinterested in exploring the interaction between CYP17 c.1-34T>C and menopausal hormones and diabetes mellitus asboth are breast cancer risk factors, possibly acting throughincreased levels of exogenous and endogenous estrogens,respectively. Thus, in order to increase statistical power in thesubgroup analyses, we further selected all remaining eligiblewomen (191 cases and 108 controls) that had used meno-pausal hormones (estrogen only or any combination of estro-gen and progestin) for at least 4 years and all women (110cases and 104 controls) with self-reported diabetes mellitus.In addition, 345 eligible controls from the founding study, orig-inally selected for a parallel endometrial cancer study but whofulfilled the inclusion criteria to the breast cancer study, wereadded to our control sample. In total, we selected 1,801 casesand 2,057 controls.

Paraffin-embedded, non-cancerous tissue samples were col-lected for deceased breast cancer cases as well as for thosebreast cancer cases that declined to donate blood but con-sented to our use of the tissue. We obtained blood samples orarchived tissue samples for 1,322 and 247 breast cancer

patients, respectively, and blood samples from 1,524 controls.Population-based participation rates (taking into accountthose that did not participate in the founding study but couldhave been selected) for cases and controls were 73% and61%, respectively.

We isolated DNA from 3 ml of whole blood using a WizardGenomic DNA Purification Kit (Promega, Madison, WI, USA)according to the manufacturer's instructions. From non-malig-nant cells in paraffin-embedded tissue, we extracted DNAusing a standard phenol/chloroform/isoamyl alcohol protocol[29]. The study was approved by the five Swedish InstitutionalReview Boards and was performed in compliance with theHelsinki Declaration.

Genetic analysesWe used two methods for CYP17 c.1-34T>C (rs743572)genotyping: Multiplex fluorescent solid-phase minisequencing[30]; and dynamic allele specific hybridization (DASH) [31].Results from the two methods were validated with PCR-RFLP[32]. Twenty-four percent of the samples were analysed withboth minisequencing and DASH, and the genotypes obtainedwere identical. Primer and probe sequences and PCR reactionconditions are given in Additional file 1. The PCR reactionswere performed on a Perkin Elmer GeneAmp 9700 system(PerkinElmer Life And Analytical Sciences, Inc., Boston, MA,USA).

Statistical analysesOdds ratios (ORs) and 95% confidence intervals (CIs) werecalculated using conditional logistic regression models withthe A1/A1 genotype as the reference group. We conditionedthe models on age (due to matching) and on duration of men-opausal hormone use and self-reported diabetes mellitus (dueto our over-sampling scheme). The conditioning variableswere thus age (5-year age groups), menopausal estrogen onlyuse (never, <4 years or ≥4 years), menopausal estrogen andprogestin use (never, <4 years or ≥4 years) and self-reporteddiabetes mellitus (yes or no). We chose to condition becausein the presence of interaction between any of the selection var-iables and CYP17 genotype, estimates of the main effect ofCYP17 could be biased. We did not expect confounding byother known breast cancer risk factors as they most likelywould be intermediates between the genetic variant andbreast cancer. We nevertheless included important breastcancer risk factors as co-variates in the models and assessedchanges in the estimates of association between CYP17 gen-otype and breast cancer to investigate if there were any indi-cations of such interrelations. Breast cancer risk factors otherthan menopausal hormone use that were analyzed for gene-environment interaction were: Age at menarche (≤2 years,>12–14 years or >14 years) and age at menopause (<49years, 49–52 years or >52 years) (where 1st categoryincluded the 1st quartile of the data, 2nd category the 2nd and3rd quartiles, and 3rd category the 4th quartile); age at first birth

Available online http://breast-cancer-research.com/content/7/6/R890

R892

(≤24 years, 25–29 years or ≥30 years); parity (nulliparous, 1child, 2 children or >2 children); and body mass index one yearprior to diagnosis (<25, 25 to <28, or ≥28). The last threewere categorized according to cut-offs that in previous studieshave been shown to be informative with regard to the varia-bles' influence on breast cancer risk. The likelihood ratio testand the Wald test statistic were used to test for interaction.We performed all analyses using the SAS system (release8.01; SAS Institute Inc., Cary, NC, USA).

ResultsWe obtained CYP17 c.1-34T>C genotypes from 1,544breast cancer cases and 1,502 controls. Among the controls,genotype frequencies were similar to previously published fre-quencies in Caucasian populations [5,9-11,20,21] and did notdeviate from Hardy-Weinberg Equilibrium (P = 0.927). Geno-

type frequencies among the cases did not differ betweenblood and tissue samples (P = 0.697).

We reproduced previously established breast cancer risk fac-tor associations in this study population [24,25,33] (Table 1,first column). The one exception was age at menarche, whichdid not appear to be associated with overall breast cancer riskin this study [25].

The CYP17 c.1-34T>C was not associated with other breastcancer risk factors among the randomly selected controls(Table 2).

We found no overall association between CYP17 c.1-34T>Cand breast cancer risk, regardless of histopathology (Table 3).This negative result was not modified by menopausal hormoneuse or diabetes mellitus (Table 3). We also did not detect any

Table 1

Breast cancer risk in relation to CYP17 genotype, stratified by breast cancer risk factors

All genotypes CYP17 genotype P for interaction

A1/A1 A1/A2 A2/A2

Cases/controlsa

ORb (CI) Cases/controlsa

ORb (CI) Cases/controlsa

ORb (CI) Cases/controlsa

ORb (CI)

Age at menarche (years)

≤12 314/262 1 (reference) 102/96 1 (reference) 138/128 1.1 (0.7–1.5) 74/38 1.9 (1.1–3.0)

>12–14 749/702 0.9 (0.7–1.1) 283/251 1.1 (0.8–1.5) 360/325 1.1 (0.8–1.5) 106/126 0.8 (0.6–1.0)

>14 300/258 1.0 (0.8–1.2) 107/94 1.1 (0.7–1.6) 152/126 1.2 (0.8–1.7) 41/38 1.1 (0.6–1.8) 0.026

Age at menopause (years)

<49 330/347 1 (reference) 145/123 1 (reference) 141/165 0.7 (0.5–1.0) 44/59 0.7 (0.4–1.1)

49–52 802/656 1.3 (1.1–1.6) 282/243 1.0 (0.7–1.3) 380/316 1.0 (0.8–1.4) 140/97 1.2 (0.9–1.8)

>52 357/323 1.2 (1.0–1.5) 119/117 1.0 (0.6–1.3) 184/150 1.1 (0.8–1.6) 54/56 0.8 (0.5–1.3) 0.062

Age at first birth (years)

≤24 616/639 1 (reference) 225/236 1 (reference) 310/315 1.1 (0.8–1.3) 81/88 1.0 (0.7–1.4)

25–29 414/388 1.1 (0.9–1.4) 145/144 1.1 (0.8–1.5) 187/171 1.2 (0.9–1.6) 82/73 1.2 (0.8–1.7)

≥30 242/187 1.4 (1.1–1.7) 100/67 1.6 (1.1–2.3) 107/89 1.3 (0.9–1.8) 35/31 1.1 (0.7–1.9) 0.752

Parity

Nulliparous 224/124 1 (reference) 79/41 1 (reference) 105/63 0.9 (0.5–1.5) 40/20 1.2 (0.6–2.3)

1 child 320/240 0.7 (0.5–0.9) 113/77 0.8 (0.5–1.2) 168/130 0.7 (0.4–1.0) 39/33 0.6 (0.3–1.1)

2 children 569/515 0.6 (0.4–0.7) 215/195 0.6 (0.4–0.9) 254/238 0.5 (0.4–0.8) 100/82 0.6 (0.4–1.0)

>2 children 386/459 0.5 (0.4–0.6) 143/175 0.4 (0.3–0.7) 184/207 0.5 (0.3–0.7) 59/77 0.4 (0.2–0.6) 0.792

Body mass index (kg/m2)

<25 716/669 1 (reference) 263/230 1 (reference) 337/329 0.9 (0.7–1.2) 116/110 0.9 (0.7–1.3)

25 to <28 371/377 1.0 (0.8–1.1) 141/137 0.9 (0.7–1.2) 170/181 0.9 (0.7–1.2) 60/59 0.9 (0.6–1.4)

≥28 403/278 145/116 1.2 (0.9–1.6) 198/121 1.4 (1.1–1.9) 60/41 1.4 (0.9–2.2) 0.657

aIncluding only cases and controls with complete information on menopausal hormones and diabetes mellitus. bAnalyses were conditioned on age (5-year age groups), menopausal estrogen only use (never, <4 years or ≥4 years), use of estrogen in combination with progestin (never, <4 years or ≥4 years) and diabetes mellitus (yes or no). CI, confidence interval; OR, odds ratio.

Breast Cancer Research Vol 7 No 6 Einarsdóttir et al.

R893

association when we considered stage 1 (n = 389) and stage2 or more advanced (n = 591) breast cancers separately (datanot shown). None of these findings were altered by restrictingthe sample set to the randomly selected cases and controls orby including other breast cancer risk factors as co-variates inthe logistic regression models.

In exploratory analyses, the A2/A2 genotype, compared to A1/A1, was associated with an increased breast cancer risk inwomen with age at menarche of 12 years or younger (OR 1.9,95% CI 1.1–3.0; P for interaction = 0.026; Table 1). Further-more, the A2 allele in carriers conferred a decreased risk inwomen with menopause before 49 years of age; OR 0.7 (95%CI 0.5–1.0) for heterozygotes and OR 0.7 (95% CI 0.4–1.1)for homozygotes) compared to A1/A1. There was, however,no dose-response pattern in these findings and we thereforeregarded them as being due to chance. Age at first birth, parityor body mass index did not seem to modify the non-associa-tion between CYP17 c.1-34T>C genotype and breast cancerrisk (Table 1). All estimates remained unaffected after werestricted the analyses to the randomly selected cases andcontrols.

DiscussionCYP17 c.1-34T>C was not associated with breast cancer riskoverall, a result that did not seem to be modified by menopau-sal hormone use or diabetes mellitus. Incidentally, we detecteda statistically significant interaction between CYP17 c.1-

34T>C and age at menarche but regarded the finding as achance result as no dose-response pattern was evident.

Our population-based study is one of the largest published todate, investigating the possible influence of CYP17 variationon breast cancer risk. Retrieval of extensive information onreproductive and lifestyle factors enabled us to test for gene-environment interactions. Misclassification of genotypesshould be minimal as is indicated by the identical results fromreplicate genotyping with two different methods, all blinded tocase-control status. Any genotype misclassification despitethese measures is unlikely to have been differential and wouldconsequently only bias our estimates towards the null.

Because pre- and postmenopausal breast cancer is con-ceived to have differing causal mechanisms, these two typesof cancer are often investigated separately. We were inter-ested in potential interaction between CYP17 genotype andmenopausal hormone treatment, which motivated our restrict-ing the study population to postmenopausal women. Althoughit would be interesting to know also if CYP17 genotype influ-ences the risk of premenopausal breast cancer, had weincluded both pre- and postmenopausal women, the study sizewould have had to be increased to have the same power toaddress our main hypotheses. We were also interested inexploring the interaction with diabetes mellitus as insulinresistance can lead to increased levels of plasma insulin, lowerlevels of plasma sex hormone binding globulin and, conse-quently, increased levels of free estrogen [34]. We thus over-

Table 2

Descriptive characteristics of postmenopausal breast cancer cases and controls by CYP17 c.1-34T>C genotype

CYP17 genotype

A1/A1 A1/A2 A2/A2

Characteristic Cases/controlsa

Cases Controls Cases/controlsa

Cases Controls Cases/controlsa

Cases Controls

Genotype frequencies 550/488 36.7 36.5 711/638 47.4 47.7 238/212 15.9 15.8

Age (years)b 550/488 63.3 ± 6.5 63.0 ± 6.5 711/638 63.4 ± 6.5 62.9 ± 6.4 238/212 62.8 ± 6.6 63.2 ± 6.1

Age at menarche (years)b 492/441 13.5 ± 1.4 13.5 ± 1.4 650/579 13.6 ± 1.4 13.5 ± 1.4 221/202 13.2 ± 1.5 13.5 ± 1.3

Age at menopause (years)b 546/483 49.9 ± 3.8 50.1 ± 4.1 705/631 50.7 ± 3.2 50.0 ± 3.8 238/212 50.6 ± 3.2 50.2 ± 4.2

Age at first birth (years)b 470/447 25.4 ± 5.0 24.7 ± 4.6 604/575 25.1 ± 5.0 24.7 ± 4.7 198/192 25.6 ± 4.5 25.2 ± 4.9

Parityb 550/488 1.9 ± 1.2 2.3 ± 1.4 711/638 1.8 ± 1.2 2.1 ± 1.2 238/212 1.8 ± 1.2 2.2 ± 1.3

Body mass index (kg/m2)b 549/483 25.8 ± 4.1 25.8 ± 4.4 705/631 25.9 ± 4.2 25.2 ± 3.8 236/210 25.7 ± 4.4 25.3 ± 4.0

Duration of menopausal hormone use (years)c

0 390/370 70.9 75.8 469/473 66.0 74.1 165/154 69.3 72.6

<4 59/50 10.7 10.3 96/74 13.5 11.6 26/25 10.9 11.8

≥4d 101/68 18.4 13.9 146/91 20.5 14.3 47/33 19.8 15.6

Self-reported diabetes mellitusc,d 550/488 8.73 8.20 711/638 9.42 7.21 238/212 8.82 8.49

aIncluding only cases and controls with complete information on menopausal hormones and diabetes mellitus. bMeans ± standard deviation. cPercentages. dLong term users of menopausal hormones and women with diabetes mellitus were over-sampled.

Available online http://breast-cancer-research.com/content/7/6/R890

R894

sampled menopausal hormone users and diabetics in thedesign phase of the study so as to increase statistical preci-sion in subgroup analyses. We explored only the effect ofCYP17 c.1-34T>C on breast cancer risk in this article, but 13variants in 7 genes were genotyped in the current phase of thestudy. Results from three of the studied genes in relation tobreast cancer risk have previously been published [35-37].

The lack of association between CYP17 c.1-34T>C and over-all breast cancer risk in our study is in line with results from 14previous studies – where 10 included only Caucasian women– and a recent meta-analysis [5,9-22]. In contrast to Feigelsonand colleagues [8], we did not find any association betweenthis polymorphism and advanced breast cancer. Two groupshave reported an association between CYP17 genotype andbreast cancer risk in postmenopausal women [6,7]. All threestudies were performed in non-Caucasian populations.

A possible mechanism for the CYP17 c.1-34T>C polymor-phism to influence breast cancer risk is through increased bio-synthesis of and, thereby, increased levels of circulatingestrogen. Contrary to the predicted effect of the A2 allele, onegroup found decreased CYP17 mRNA levels in A2 carriers[38]. Studies regarding association of CYP17 c.1-34T>C andcirculating hormone levels as well as markers of hormonal sta-tus (i.e. age at menarche or menopausal hormone use) haverecently been reviewed [39]. Increased estradiol levels havebeen associated with the A2 allele in premenopausal women[6,32] as well as in postmenopausal women [9,40], but threegroups did not report any significant changes in hormone lev-els by genotype, two in postmenopausal women [17,41] andone in premenopausal women [42]. Results from seven stud-ies have indicated a moderate association between the A2allele and earlier menarche [8,11,15,18,43-45], an associa-tion that was not detected in our study, nor in five others

Table 3

Breast cancer risk in relation to CYP17 genotype, stratified by histopathological type, menopausal hormones or diabetes mellitus

CYP17 genotype P for interaction

A1/A1 A1/A2 A2/A2

Cases/controlsa ORb(CI) Cases/controlsa ORb(CI) Cases/controlsa ORb(CI)

All cancers 550/488 1 (reference) 711/638 1.0 (0.9–1.2) 238/212 1.0 (0.8–1.3)

Ductal cancers 420/488 1 (reference) 510/638 1.0 (0.8–1.1) 180/212 1.0 (0.8–1.3)

Lobular cancers 56/488 1 (reference) 90/638 1.3 (0.9–1.8) 24/212 1.1 (0.6–1.8)

Duration of menopausal hormone use(all cancers)

Never 390/370 1 (reference) 469/473 0.9 (0.8–1.1) 165/154 1.0 (0.8–1.3)

Any kind

<4 59/50 1 (reference) 96/74 1.2 (0.7–2.0) 26/25 1.0 (0.5–2.1)

≥4 101/68 1 (reference) 146/91 1.2 (0.8–1.9) 47/33 1.0 (0.6–1.7) 0.668

Estrogen only

<4 28/22 1 (reference) 43/33 1.1 (0.5–2.3) 7/11 0.6 (0.2–1.8)

≥4 39/23 1 (reference) 42/30 1.0 (0.5–2.1) 15/13 0.7 (0.3–1.9) 0.703

Estrogen + progestin

<4 42/41 1 (reference) 75/66 1.2 (0.7–2.1) 23/21 1.2 (0.6–2.5)

≥4 64/47 1 (reference) 108/62 1.3 (0.8–2.2) 34/20 1.2 (0.6–2.4) 0.723

Self-reported diabetes mellitus (all cancers)

No 502/448 1 (reference) 644/592 1.0 (0.8–1.2) 217/194 1.0 (0.8–1.3)

Yes 48/40 1 (reference) 67/46 1.3 (0.7–2.2) 21/18 1.1 (0.5–2.3) 0.259

aIncluding only cases and controls with complete information on menopausal hormones and diabetes mellitus. bAnalyses were conditioned on age (5-year age groups), menopausal estrogen only use (never, <4 years or ≥4 years), use of estrogen in combination with progestin (never, <4 years or ≥4 years) and diabetes mellitus (yes or no). CI, confidence interval; OR, odds ratio.

Breast Cancer Research Vol 7 No 6 Einarsdóttir et al.

R895

[9,12,21,46,47]. Furthermore, previous investigators haveposited that CYP17 genotype may be associated with use ofmenopausal hormones, an important risk factor for breast can-cer, but results have been inconsistent [11,12,21,48-50]. Wefound no such association in our data.

We identified an interaction with age at menarche, but consid-ered it unlikely that the A2/A2 genotype would increase breastcancer risk in women with age at menarche less than 13 yearswithout also moderately increasing risk in women with age atmenarche between 12 and 14 years. Instead, the A2/A2 gen-otype decreased risk in the latter group, which is difficult toexplain biologically. A similar pattern was seen for the interac-tion with age at menopause. To our knowledge, other groupshave not reported comparable findings and we thereforebelieve the results to be caused by chance alone.

ConclusionCYP17 c.1-34T>C does not appear to have any major influ-ence on breast cancer risk. Despite this negative finding, itwould be prudent to investigate common variants covering theentire CYP17 gene before concluding that this gene has norole in breast cancer etiology.

Competing interestsThe authors declare that they have no competing interests.

Authors' contributionsKristjana Einarsdóttir wrote the article and performed the sta-tistical analyses. Keith Humphreys and Gudrun Jonasdottiradvised on the statistical methods. Tove Rylander-Rudqvistand Susanne Ahlberg performed the genetic analyses. Mag-nus Ingelman-Sundberg and Ingemar Persson conceived anddesigned the study. Elisabete Weiderpass oversaw theproject progress and took part in interpretation and writing thearticle. Jianjun Liu and Kee Seng Chia critically revised themanuscript. Per Hall was involved in drafting the article,assisted in interpretation of data, and gave final approval of theversion to be published. Sara Wedrén contributed to concep-tion and design of the project, coordinated sample collection,was involved in writing the article, assisted in analysis andinterpretation of data and gave final approval for publication.All authors read and approved the final manuscript.

Additional files

AcknowledgementsWe thank all the women who participated in the study and made it pos-sible, especially those who took the effort to donate a blood sample. We are also grateful to Anthony S Gunnell for editing assistance and Fredrik Granath for statistical advice. This work was funded by the National Insti-tute of Health, grant number R03 CA83114, the Swedish Cancer Soci-ety, grant number 4112-B99-02XBB, and the K&A Wallenberg foundation (Wallenberg Consortium North, K Humphreys).

References1. Bernstein L: Epidemiology of endocrine-related risk factors for

breast cancer. J Mammary Gland Biol Neoplasia 2002, 7:3-15.2. Nakajin S, Shinoda M, Haniu M, Shively JE, Hall PF: C21 steroid

side chain cleavage enzyme from porcine adrenal micro-somes. Purification and characterization of the 17 alpha-hydroxylase/C17,20-lyase cytochrome P-450. J Biol Chem1984, 259:3971-3976.

3. Picado-Leonard J, Miller WL: Cloning and sequence of thehuman gene for P450c17 (steroid 17 alpha-hydroxylase/17,20lyase): similarity with the gene for P450c21. DNA 1987,6:439-448.

4. Carey AH, Waterworth D, Patel K, White D, Little J, Novelli P,Franks S, Williamson R: Polycystic ovaries and premature malepattern baldness are associated with one allele of the steroidmetabolism gene CYP17. Hum Mol Genet 1994, 3:1873-1876.

5. Nedelcheva Kristensen V, Haraldsen EK, Anderson KB, LonningPE, Erikstein B, Karesen R, Gabrielsen OS, Borresen-Dale AL:CYP17 and breast cancer risk: the polymorphism in the 5'flanking area of the gene does not influence binding to Sp-1.Cancer Res 1999, 59:2825-2828.

6. Chacko P, Rajan B, Mathew BS, Joseph T, Pillai MR: CYP17 andSULT1A1 gene polymorphisms in Indian breast cancer. BreastCancer 2004, 11:380-388.

7. Miyoshi Y, Iwao K, Ikeda N, Egawa C, Noguchi S: Genetic poly-morphism in CYP17 and breast cancer risk in Japanesewomen. Eur J Cancer 2000, 36:2375-2379.

8. Feigelson HS, Coetzee GA, Kolonel LN, Ross RK, Henderson BE:A polymorphism in the CYP17 gene increases the risk ofbreast cancer. Cancer Res 1997, 57:1063-1065.

9. Haiman CA, Hankinson SE, Spiegelman D, Colditz GA, WillettWC, Speizer FE, Kelsey KT, Hunter DJ: The relationship betweena polymorphism in CYP17 with plasma hormone levels andbreast cancer. Cancer Res 1999, 59:1015-1020.

10. Mitrunen K, Jourenkova N, Kataja V, Eskelinen M, Kosma VM, Ben-hamou S, Vainio H, Uusitupa M, Hirvonen A: Steroid metabolismgene CYP17 polymorphism and the development of breastcancer. Cancer Epidemiol Biomarkers Prev 2000, 9:1343-1348.

11. Ambrosone CB, Moysich KB, Furberg H, Freudenheim JL, Bow-man ED, Ahmed S, Graham S, Vena JE, Shields PG: CYP17genetic polymorphism, breast cancer, and breast cancer riskfactors. Breast Cancer Res 2003, 5:R45-51.

12. Helzlsouer KJ, Huang HY, Strickland PT, Hoffman S, Alberg AJ,Comstock GW, Bell DA: Association between CYP17 polymor-phisms and the development of breast cancer. Cancer Epide-miol Biomarkers Prev 1998, 7:945-949.

13. Huang CS, Chern HD, Chang KJ, Cheng CW, Hsu SM, Shen CY:Breast cancer risk associated with genotype polymorphism ofthe estrogen-metabolizing genes CYP17, CYP1A1, and COMT:a multigenic study on cancer susceptibility. Cancer Res 1999,59:4870-4875.

14. Gudmundsdottir K, Thorlacius S, Jonasson JG, Sigfusson BF, Try-ggvadottir L, Eyfjord JE: CYP17 promoter polymorphism andbreast cancer risk in males and females in relation to BRCA2status. Br J Cancer 2003, 88:933-936.

15. Weston A, Pan CF, Bleiweiss IJ, Ksieski HB, Roy N, Maloney N,Wolff MS: CYP17 genotype and breast cancer risk. Cancer Epi-demiol Biomarkers Prev 1998, 7:941-944.

16. Hefler LA, Tempfer CB, Grimm C, Lebrecht A, Ulbrich E, Heinze G,Leodolter S, Schneeberger C, Mueller MW, Muendlein A, et al.:Estrogen-metabolizing gene polymorphisms in the assess-ment of breast carcinoma risk and fibroadenoma risk in Cau-casian women. Cancer 2004, 101:264-269.

17. Dunning AM, Dowsett M, Healey CS, Tee L, Luben RN, Folkerd E,Novik KL, Kelemen L, Ogata S, Pharoah PD, et al.: Polymor-

The following Additional files are available online:

Additional File 1Microsoft Word document detailing the PCR reaction conditions as well as primer and probe sequences used to perform the minisequencing and DASH methods.See http://www.biomedcentral.com/content/supplementary/bcr1319-S1.doc

Available online http://breast-cancer-research.com/content/7/6/R890

R896

phisms associated with circulating sex hormone levels inpostmenopausal women. J Natl Cancer Inst 2004, 96:936-945.

18. Wu AH, Seow A, Arakawa K, Van Den Berg D, Lee HP, Yu MC:HSD17B1 and CYP17 polymorphisms and breast cancer riskamong Chinese women in Singapore. Int J Cancer 2003,104:450-457.

19. Ahsan H, Whittemore AS, Chen Y, Senie RT, Hamilton SP, WangQ, Gurvich I, Santella RM: Variants in estrogen-biosynthesisgenes CYP17 and CYP19 and breast cancer risk: a family-based genetic association study. Breast Cancer Res 2005,7:R71-81.

20. Verla-Tebit E, Wang-Gohrke S, Chang-Claude J: CYP17 5'-UTRMspA1 polymorphism and the risk of premenopausal breastcancer in a German population-based case-control study.Breast Cancer Res 2005, 7:R455-464.

21. Chang JH, Gertig DM, Chen X, Dite GS, Jenkins MA, Milne RL,Southey MC, McCredie MR, Giles GG, Chenevix-Trench G, et al.:CYP17 genetic polymorphism, breast cancer, and breast can-cer risk factors: Australian Breast Cancer Family Study. BreastCancer Res 2005, 7:R513-521.

22. Ye Z, Parry JM: The CYP17 MspA1 polymorphism and breastcancer risk: a meta-analysis. Mutagenesis 2002, 17:119-126.

23. Feigelson HS, McKean-Cowdin R, Henderson BE: Concerningthe CYP17 MspA1 polymorphism and breast cancer risk: ameta-analysis. Mutagenesis 2002, 17:445-446. author reply447–448

24. Magnusson C, Colditz G, Rosner B, Bergstrom R, Persson I:Association of family history and other risk factors with breastcancer risk (Sweden). Cancer Causes Control 1998,9:259-267.

25. Magnusson CM, Persson IR, Baron JA, Ekbom A, Bergstrom R,Adami HO: The role of reproductive factors and use of oralcontraceptives in the aetiology of breast cancer in womenaged 50 to 74 years. Int J Cancer 1999, 80:231-236.

26. Magnusson C, Baron JA, Correia N, Bergstrom R, Adami HO,Persson I: Breast-cancer risk following long-term oestrogen-and oestrogen-progestin-replacement therapy. Int J Cancer1999, 81:339-344.

27. Moradi T, Nyren O, Zack M, Magnusson C, Persson I, Adami HO:Breast cancer risk and lifetime leisure-time and occupationalphysical activity (Sweden). Cancer Causes Control 2000,11:523-531.

28. Terry P, Wolk A, Persson I, Magnusson C: Brassica vegetablesand breast cancer risk. JAMA 2001, 285:2975-2977.

29. Isola J, DeVries S, Chu L, Ghazvini S, Waldman F: Analysis ofchanges in DNA sequence copy number by comparativegenomic hybridization in archival paraffin-embedded tumorsamples. Am J Pathol 1994, 145:1301-1308.

30. Pastinen T, Partanen J, Syvanen AC: Multiplex, fluorescent,solid-phase minisequencing for efficient screening of DNAsequence variation. Clin Chem 1996, 42:1391-1397.

31. Prince JA, Feuk L, Howell WM, Jobs M, Emahazion T, Blennow K,Brookes AJ: Robust and accurate single nucleotide polymor-phism genotyping by dynamic allele-specific hybridization(DASH): design criteria and assay validation. Genome Res2001, 11:152-162.

32. Feigelson HS, Shames LS, Pike MC, Coetzee GA, Stanczyk FZ,Henderson BE: Cytochrome P450c17alpha gene (CYP17) pol-ymorphism is associated with serum estrogen and progester-one concentrations. Cancer Res 1998, 58:585-587.

33. Magnusson C, Baron J, Persson I, Wolk A, Bergstrom R, Tri-chopoulos D, Adami HO: Body size in different periods of lifeand breast cancer risk in post-menopausal women. Int JCancer 1998, 76:29-34.

34. Kaaks R: Nutrition, hormones, and breast cancer: is insulin themissing link? Cancer Causes Control 1996, 7:605-625.

35. Wedren S, Rudqvist TR, Granath F, Weiderpass E, Ingelman-Sun-dberg M, Persson I, Magnusson C: Catechol-O-methyltrans-ferase gene polymorphism and post-menopausal breastcancer risk. Carcinogenesis 2003, 24:681-687.

36. Rylander-Rudqvist T, Wedren S, Granath F, Humphreys K, AhlbergS, Weiderpass E, Oscarson M, Ingelman-Sundberg M, Persson I:Cytochrome P450 1B1 gene polymorphisms and postmeno-pausal breast cancer risk. Carcinogenesis 2003,24:1533-1539.

37. Wedren S, Lovmar L, Humphreys K, Magnusson C, Melhus H,Syvanen AC, Kindmark A, Landegren U, Fermer ML, Stiger F, et al.:

Oestrogen receptor alpha gene haplotype and postmenopau-sal breast cancer risk: a case control study. Breast Cancer Res2004, 6:R437-449.

38. Daneshmand S, Weitsman SR, Navab A, Jakimiuk AJ, MagoffinDA: Overexpression of theca-cell messenger RNA in poly-cystic ovary syndrome does not correlate with polymorphismsin the cholesterol side-chain cleavage and 17alpha-hydroxy-lase/C(17–20) lyase promoters. Fertil Steril 2002, 77:274-280.

39. Sharp L, Cardy AH, Cotton SC, Little J: CYP17 gene polymor-phisms: prevalence and associations with hormone levels andrelated factors. a HuGE review. Am J Epidemiol 2004,160:729-740.

40. Hong CC, Thompson HJ, Jiang C, Hammond GL, Tritchler D, YaffeM, Boyd NF: Association between the T27C polymorphism inthe cytochrome P450 c17alpha (CYP17) gene and risk factorsfor breast cancer. Breast Cancer Res Treat 2004, 88:217-230.

41. Tworoger SS, Chubak J, Aiello EJ, Ulrich CM, Atkinson C, PotterJD, Yasui Y, Stapleton PL, Lampe JW, Farin FM, et al.: Associationof CYP17, CYP19, CYP1B1, and COMT polymorphisms withserum and urinary sex hormone concentrations in postmeno-pausal women. Cancer Epidemiol Biomarkers Prev 2004,13:94-101.

42. Garcia-Closas M, Herbstman J, Schiffman M, Glass A, Dorgan JF:Relationship between serum hormone concentrations, repro-ductive history, alcohol consumption and genetic polymor-phisms in pre-menopausal women. Int J Cancer 2002,102:172-178.

43. Hamajima N, Iwata H, Obata Y, Matsuo K, Mizutani M, Iwase T,Miura S, Okuma K, Ohashi K, Tajima K: No association of the 5'promoter region polymorphism of CYP17 with breast cancerrisk in Japan. Jpn J Cancer Res 2000, 91:880-885.

44. Gorai I, Tanaka K, Inada M, Morinaga H, Uchiyama Y, Kikuchi R,Chaki O, Hirahara F: Estrogen-metabolizing gene polymor-phisms, but not estrogen receptor-alpha gene polymor-phisms, are associated with the onset of menarche in healthypostmenopausal Japanese women. J Clin Endocrinol Metab2003, 88:799-803.

45. Lai J, Vesprini D, Chu W, Jernstrom H, Narod SA: CYP gene pol-ymorphisms and early menarche. Mol Genet Metab 2001,74:449-457.

46. Dunning AM, Healey CS, Pharoah PD, Foster NA, Lipscombe JM,Redman KL, Easton DF, Day NE, Ponder BA: No associationbetween a polymorphism in the steroid metabolism geneCYP17 and risk of breast cancer. Br J Cancer 1998,77:2045-2047.

47. Goodman MT, McDuffie K, Guo C, Terada K, Donlon TA: CYP17genotype and ovarian cancer: a null case-control study. Can-cer Epidemiol Biomarkers Prev 2001, 10:563-564.

48. Feigelson HS, McKean-Cowdin R, Pike MC, Coetzee GA, KolonelLN, Nomura AM, Le Marchand L, Henderson BE: CytochromeP450c17alpha gene (CYP17) polymorphism predicts use ofhormone replacement therapy. Cancer Res 1999,59:3908-3910.

49. Haiman CA, Hankinson SE, Colditz GA, Hunter DJ, De Vivo I: Apolymorphism in CYP17 and endometrial cancer risk. CancerRes 2001, 61:3955-3960.

50. McKean-Cowdin R, Feigelson HS, Pike MC, Coetzee GA, KolonelLN, Henderson BE: Risk of endometrial cancer and estrogenreplacement therapy history by CYP17 genotype. Cancer Res2001, 61:848-849.